challenge: end of life for 6.2 is premature with buggy 6.3

[Paul Schmehl]

--On Thursday, June 05, 2008 14:22:00 -0400 John Baldwin <jhb at freebsd.org> 
wrote:
>
> I find that bce(4) is far more reliable in 6.3 than 6.1 for us.  There have
> been several fixes (esp. for higher loads, and mostly in 6.2) to this driver.
> There are known panics in earlier 6.x that are fixed in 6.3 for certain with
> this driver.
>

Thanks.  Knowing that gives me a lot more confidence to go ahead and build a 
new kernel for that server.

> In general though, you don't know which bugs are fixed and if any regressions
> are present w/o testing the code.  If you have production systems then
> hopefully you have QA systems for development, etc. and you can either reuse
> those when app QA isn't active for OS QA or you can get dedicated boxes for
> OS QA.  Even if you used a commercial OS with a support contract you would
> need to do the same.

Again, that would be nice, but **just like FreeBSD** this is an all volunteer 
project where both time and money are at a premium.  If I had a dollar for 
every time my wife complained about me using my valuable free time to support 
this site without any compensation, I could probably afford a test bed. :-)

-- 
Paul Schmehl
As if it wasn't already obvious,
my opinions are my own and not
those of my employer.